In recent years, there has been a growing concern about the issue of racial and gender bias in law enforcement, and Dallas, Texas, has not been immune to this trend. The city has witnessed several incidents of police misconduct, excessive force, and racial targeting over the past decade. One such incident occurred in 2018, when an unarmed Black man, Botham Jean, was fatally shot by a white police officer in his own apartment, sparking nationwide protests. Another incident involved a female police officer who was fired in 2020 for making derogatory comments about transgender individuals on social media, highlighting the issue of gender bias in law enforcement.
At the same time, a report on the Dallas police-community relations panel in 2021 revealed that more work needs to be done to address issues of diversity and power in the panel. These incidents have raised important questions about the culture and practices of law enforcement in Dallas, and the need for reforms to eliminate racial and gender biases.
The purpose of this project is to analyze and visualize the findings on police use of force incidents in Dallas, Texas. The project aimed to identify any biases or patterns that may exist in the data, particularly concerning the subject’s race and gender, and provide recommendations for improving police use of force practices. Through the analysis of data for the year 2016, various data visualization techniques were used to create visualizations that provide insights into patterns and trends in the data, such as the days of the incident, type of force used, incidents by month, injured officers, and injured subjects.
The table below revealed that male officers were involved in a significantly higher number of incidents overall, with 1,772 incidents involving male officers and male subjects. However, there were also incidents involving female officers, with a total of 77 incidents involving female officers and female subjects and 160 incidents involving female officers and male subjects.
Table 1: Incidents by Subject Race
| Female | Male | NULL | Unknown | |
|---|---|---|---|---|
| Female | 77 | 160 | 2 | 1 |
| Male | 363 | 1772 | 8 | 0 |
Figure 1: Distribution of Incidents by gender
The resulting plot below suggests clearly that there may be gender-specific patterns in police incidents.
Table 2 shows that the majority of police incidents (55.94%) involved black subjects, followed by Hispanic subjects at 21.99% and white subjects at 19.72%. Other races and unknown races accounted for a small percentage of incidents. Table 3 shows that white officers were the most frequent participants in incidents, accounting for nearly 62% of all cases. Hispanic officers accounted for 20.23% of incidents, followed by black officers at 14.31%. Other races and unknown races accounted for a small percentage of incidents.
Table 2: Officer involved in incident by Race
| OFFICER_RACE | count | percentage |
|---|---|---|
| White | 1470 | 61.69 |
| Hispanic | 482 | 20.23 |
| Black | 341 | 14.31 |
| Asian | 55 | 2.31 |
| Other | 27 | 1.13 |
| American Ind | 8 | 0.34 |
Table 3: Subject involved in incident by Race
| SUBJECT_RACE | count | percentage |
|---|---|---|
| Black | 1333 | 55.94 |
| Hispanic | 524 | 21.99 |
| White | 470 | 19.72 |
| NULL | 39 | 1.64 |
| Other | 11 | 0.46 |
| Asian | 5 | 0.21 |
| American Ind | 1 | 0.04 |
Figure 2 & 3: Incidents involvement by Race
The plots suggest that there may be racial disparities in police incidents, as black and Hispanic subjects are disproportionately involved, while white officers are over-represented as participants.
As shown in table 4, the frequency of incidents involving officers and subjects of different races. The majority of incidents involved white officers and black subjects, for a total of 846 incidents. This was followed by incidents involving Hispanic officers and black subjects, for a total of 230 incidents. Incidents involving American Indian officers had the lowest count among all officer races, with only 8 incidents reported in total.
Table 4: Frequency of Incidents by Officer and Subject Race
| American Ind | Asian | Black | Hispanic | NULL | Other | White | |
|---|---|---|---|---|---|---|---|
| American Ind | 0 | 0 | 7 | 0 | 0 | 0 | 1 |
| Asian | 0 | 0 | 28 | 9 | 1 | 0 | 17 |
| Black | 0 | 1 | 201 | 73 | 9 | 0 | 57 |
| Hispanic | 1 | 0 | 230 | 138 | 7 | 2 | 104 |
| Other | 0 | 0 | 21 | 2 | 0 | 0 | 4 |
| White | 0 | 4 | 846 | 302 | 22 | 9 | 287 |
The figure 4 below shows the frequency of incidents by officer and subject race. It highlights the large number of incidents involving white officers and black subjects.
Figure 4: Frequency of Incidents by Officer and Subject Race
The findings of the race analysis suggest that incidents involving black subjects were the most frequent, and white officers were the most frequent participants in incidents. This provides valuable insights into the relationship between the race of police officers and the subjects involved in incidents. These findings raise concerns about potential racial disparities in police incidents and highlight the need for further investigation and efforts to address any biases or discrimination in law enforcement.
The plotted data clearly shows difference in the number of incidents reported on various days of the week. Incidents on Sundays and Saturdays were reported more frequently than any other day, with Friday having the third-highest number of reported incidents. Thursday, Tuesday, and Wednesday had a relatively lower number of incidents reported compared to the other days. Interestingly, Monday had the lowest number of incidents reported among all the days of the week.
Table 4: Incidents by days of the week
| INCIDENT_WEEKDAYS | n |
|---|---|
| Friday | 379 |
| Monday | 274 |
| Saturday | 393 |
| Sunday | 428 |
| Thursday | 313 |
| Tuesday | 310 |
| Wednesday | 286 |
Figure 5: Incidents by days of the week
Based on our analysis, we found that there were more incidents reported on weekends, particularly on Sundays and Saturdays.
The plot shows the trend of the incident count over time, with each point representing a month.
Figure 6: Incidents by months of the year
## Warning: 10 failed to parse.
The data shows that there was a noticeable fluctuation in the incident count over the course of the year. In particular, there was a sharp increase in March 2016, followed by a steep decline in December 2016. This suggests that there may be seasonal patterns in the incident rate.
It can be seen in the table that the most common types of force used by police officers were verbal commands, weapon displays at the person, and holding a suspect down, in that order. These three types of force were used significantly more often than other methods, including taser use and joint locks.
Table 5: Type of force used
## Selecting by n
| TYPE_OF_FORCE_USED1 | n |
|---|---|
| Verbal Command | 818 |
| Weapon display at Person | 329 |
| Held Suspect Down | 176 |
| BD - Grabbed | 154 |
| Take Down - Arm | 144 |
| Joint Locks | 140 |
| Hand Controlled Escort | 107 |
| Taser | 78 |
| Taser Display at Person | 69 |
| Take Down - Body | 68 |
Figure 7: Type of force used
Based on the plot presented in figure 7, it can be concluded that police officers prioritize using communication skills and their authority to de-escalate situations and gain control over suspects. The data also shows that the next most common methods of force used were displaying a weapon and holding a suspect down, indicating that officers tend to rely on non-lethal approaches to subdue suspects.
Based on the map, it can be inferred that incidents occurred throughout the city and were not limited to a specific neighborhood or region. This indicates that the incidents were not concentrated in any one area but rather spread out across the entire city.
Figure 8: Incidents location on map
From the plot analysis, we can conclude that there is a negative correlation between the number of incidents and years on force. This means that officers with more years of experience tend to be involved in fewer incidents. Additionally, the majority of officers have not been injured, and those who have been injured have lower incident counts. Notably, officers with 1-3 years of experience have a higher likelihood of being injured compared to those with more years of experience.
Figure 9: Year in force and injury distribution
These findings suggest that more experienced officers are better equipped to handle incidents and avoid injury, while newer officers may require additional support and training to reduce their risk of injury.
The analysis of the policing dataset from Dallas, has highlighted the need for reform in the police department to eliminate any racial or gender bias that may exist. It is crucial to ensure that police officers are trained to handle situations without resorting to excessive use of force. The use of body cameras can also be helpful in ensuring transparency and accountability.
In conclusion, the analysis of data related to police use of force incidents in Dallas, Texas, for the year 2016 has provided valuable insights into patterns and trends in the data. The use of various data visualization techniques has allowed for a clear understanding of the incidents’ characteristics and implications for the community. These findings underscore the importance of addressing any racial or gender biases in policing and the need for comprehensive reform in the police department.
Wilonsky, R. (2019, September 5). The trial of Amber Guyger, in photos. The Dallas Morning News. https://www.dallasnews.com/news/2019/09/05/the-trial-of-amber-guyger-in-photos/
Rocha, V. (2020, December 16). Fired Dallas cop’s history of disciplinary issues went unchecked for years. The Dallas Morning News. https://www.dallasnews.com/news/crime/2020/12/16/fired-dallas-cops-history-of-disciplinary-issues-went-unchecked-for-years/
Data Visualization with R. https://rkabacoff.github.io/datavis/
Flores, J. (2021, April 20). Report: Dallas police, community relations panel needs more power, diverse membership. The Dallas Morning News. https://www.dallasnews.com/news/2021/04/20/report-dallas-police-community-relations-panel-needs-more-power-diverse-membership/
Wickham, H. (2016). ggplot2: Elegant graphics for data analysis. Springer. - https://link.springer.com/book/10.1007/978-0-387-98141-3
City of Dallas. (2016). Police Use of Force Incidents. Data.gov. https://catalog.data.gov/dataset/police-use-of-force-incidents-2016
Dallas Police Department. (2019). Dallas Police Department Use of Force Report. Dallas Police Department. https://www.dallaspolice.net/Portals/0/SIB/Use%20of%20Force%20Report%202016.pdf